Projects Achievement

Funded by the European Union

This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement N° 101070192

Library of 3D Models

One of the core milestones of the SENSO3D project is the development of an extensive, high-quality 3D model repository optimized for AR/VR environments.

1300+ 3D models created or integrated, including:

Office furniture (chairs, desks, whiteboards)
Business technology (monitors, microphones, projectors)
Architectural components (walls, floors, ceilings, windows)
Decorative elements (plants, lighting fixtures, paintings)

Structured Repository with Metadata

Every object is tagged with detailed metadata including.
- Object category & material
- Texture resolution
- Lighting/shadow attributes
- Unity compatibility data
- Collision and physics settings
All assets are version-controlled and stored in a Git-based repository for easy integration and updates.

Visual Suggestion: Include a screenshot or infographic of the model repository interface with metadata fields.

Advanced AI Tools

SENSO3D leverages powerful AI-driven workflows to generate and manage 3D content efficiently:

2D-to-3D Object Reconstruction

AI models convert 2D images into fully-fledged 3D assets.
Works with both real-world photos and user-submitted sketches.
Real-time object detection from images, even with occlusions.

Texture Generation

No-texture models are now enhanced via AI-based description-to-texture generation.
A two-step process:
- AI model describes object characteristics (color, material).
- A texture synthesis engine maps textures accordingly.

Object Description & Classification

Automatic assignment of:
- Object type, function, and environment context
- Orientation and size estimation
- Keywords for prompt-based search and retrieval

Visual Suggestion: Include before/after of a 2D image → 3D model with texture + screenshot of AI classification interface.

Demo Environments: Lobby & Conference Room

To showcase the real-world usability of the system, two fully interactive and modular demo environments were developed in Unity:

3D Lobby

A welcoming virtual entrance environment with:
- Reception desk, seating, plants, screens
- Reception desk, seating, plants, screens

3D Conference Room

Designed for virtual meetings and collaboration:
- Includes configurable tables, ergonomic chairs, whiteboards, and AV tech
- Supports realistic physics and spatial audio

Scenes are

Built from reusable prefabs
Populated using AI-generated models
Compatible with Unity XR and WebXR

Visual Suggestion: Add panoramic screenshots of both environments showing lighting, layout, and interaction.

Prompt-Based Scene Creation Tool

One of the flagship innovations of Sprint 3 is the Prompt-Based Tool, enabling users to create entire 3D environments using natural language.

How it works:

User describes a scene (e.g., “A modern meeting room with a round table and eight chairs”).
The AI interprets the prompt using NLP (Natural Language Processing).
It fetches relevant 3D models from the repository and places them intelligently in a Unity scene.
Users can fine-tune the scene (object size, lighting, textures).

This tool dramatically lowers the entry barrier for non-technical users and accelerates prototyping.

Visual Suggestion: Show a side-by-side of the text input and the generated scene output.

KPI Summary: Goals Exceeded

SENSO3D has met or exceeded most of its Key Performance Indicators (KPIs), demonstrating strong alignment with its original objectives:

KPI	Target	Status
3D Object Library	1000+ models	1300+ achieved
AI Object Detection	50 categories	90+ with >95% accuracy
Use Case Scenarios	Virtual Meetings + Interior Design	Pilots in 3 scenarios
Prompt-Based Tool	Initial Prototype	Functional demo delivered
Unity Integration	Lobby & Conference Room	Optimized and interactive
2D Image Dataset	1000 images with segmentation	Achieved 100% target